Posts: 44
Threads: 8
Joined: Oct 2026
Reputation:
0
Posting this in case it saves someone else a little time.
I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about MoE serving.
Would rather learn from your scars than collect my own all at once.
Posts: 72
Threads: 7
Joined: Sep 2026
Reputation:
0
I'd approach it in a boring order. On 7900 XTX, the easiest win is often to back off the aggressive settings and see whether the problem survives. Once you know the plain path works, then add the faster path back in piece by piece. I keep promising myself this is the last rebuild.
Posts: 142
Threads: 20
Joined: Sep 2026
Reputation:
0
A couple questions before I guess too hard. What model, what runtime, and roughly what context size?
small models can still be useful
Posts: 211
Threads: 21
Joined: Sep 2026
Reputation:
0
Interesting, I hadn't tried that. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
7900 XTX / ROCm / somehow still optimistic
Posts: 36
Threads: 5
Joined: Sep 2026
Reputation:
0
Good point. I don't think your conclusion is far off.
Posts: 128
Threads: 12
Joined: Sep 2026
Reputation:
0
that fixed it for me too.
rack in the garage • local models everywhere
Posts: 69
Threads: 3
Joined: Sep 2026
Reputation:
0
A couple questions before I guess too hard. Are you testing this on B200 or something else?
still benchmarking everything
Posts: 24
Threads: 0
Joined: Aug 2026
Reputation:
0
Ran a quick check on my box. On GB200 with SGLang, I saw basically the same behavior. The main thing that moved was latency, not quality.